Hospital readmissions are common, costly, and potentially preventable. They are also potentially responsive to health system interventions. However, it is uncertain which components of care transition interventions are efficacious, for which populations, and at what cost. The proposed study will evaluate a three-tiered quality improvement (QI) intervention intended to reduce hospital readmissions within 30 days post-discharge from an urban safety net hospital that serves a racially and linguistically diverse population. Few studies have evaluated care transition interventions to reduce readmissions among low-income, diverse patient populations, and the accumulated evidence on the effects of these multi-faceted interventions on readmission rates has been inconclusive. This project will take advantage of a unique sequence of three QI innovations to reduce hospital readmissions implemented beginning in 2007 in an integrated safety net health care system. We propose to evaluate the additive impact of each intervention on reductions in readmissions and on post- discharge care. The discharge-transfer intervention tiers are as follows: 1) Tier 1 includes a comprehensive, individualized home care plan (HCP) reviewed by the medical service floor nurse with the patient prior to discharge; 2) Tier 2 adds the electronic transmission of the HCP to the patient's primary care medical home where, on the business day following discharge, a Registered Nurse makes an outreach telephone call to the discharged patient to confirm comprehension of the HCP and to address medical questions or needs; 3) Tier 3 further adds a community health worker, the Patient Navigator, to participate in bedside discussions to develop rapport and learn about patients' home situations, weekly outreach calls to assess patients' needs and to facilitate communication between the patient and the primary care team, and reminder calls to patients prior to all medical appointments to eliminate barriers to outpatient follow-up. Our project includes three research aims. First, in a retrospective time series analysis, we will assess the incremental effect on readmissions and health care use of intervention tiers 1 and 2 relative to the prior standard of care. Second, we will evaluate the effects of an ongoing randomized natural experiment on readmissions, health care use, adherence to medication instructions, and preparedness for discharge. This natural experiment features random assignment to one of two QI interventions, Tier 2 or Tier 3, and exclusively targets patients at high risk for readmission, those with one or more of the following risk factors for readmission: discharge diagnosis of congestive heart failure, acute coronary syndrome, or pneumonia; length of stay > 3 days; weekend discharge; age > 60; or previous hospitalization within the past six months. For our third aim, we will analyze the costs and cost-effectiveness of the Patient Navigator intervention (Tier 3) relative to the Tier 2 intervention for high-risk patients.